|
Hi everyone,
The situation:
I am comparing infection outcomes (categorical yes/no: 1/0) in 2 populations with different methods of suture (categorical: staples being 1 and another material being 0).
The table looks like this
Infection No infection
Staples 28 55
Other material 68 250
So the odds ratio (staples versus other material) for an infection outcome would be:
(28/55) / (68/250) = 0.509 / 0.272 = 1.87
And the relative risk would be
(28 / (28 + 55)) / (68/(68+250)) = (28/83) / (68/318) = 0.337 / 0.213 = 1.58.
When I enter these data into SPSS, I have 2 categorical variables;
VAR 1: suture method: 0 = other suture method, 1 = staples
VAR 2: infection: 0 = no, 1 = yes.
Creating a crosstab including with risk option enabled, entering VAR 1 in the rows and VAR 2 in the columns, gives me the following results:
RISK ESTIMATE
Odds ratio for infection (0/1) = 1.872 == so this one matches the manual calculation...
However, directly beneath it, it says (these should be the relative risks according to various sources on the web)
For cohort suture method 0 = 1.157
For cohort suture method 1 = .618.
The .618 bothers me. Because in the staple group, the infection percentage is 33.7 and in the other group this is 21.3.
So 33.7 / 21.3 should end up at 1.58 instead of 0.618
I am assuming the .618 is the relative risk for an infection comparing staples versus the other group...
Does anyone have any idea why this one doesn't match my calculation?
My goal BTW is to calculate the 95% confidence interval for the relative risk.
I don't see an option to add an attachment, but I can send the *.sav file by email, please let me know.
Help is very appreciated.
Thanks
Reaver
|
Login
Register
FAQ
Search
